diff --git a/vllm/model_executor/layers/fused_moe/gpt_oss_triton_kernels_moe.py b/vllm/model_executor/layers/fused_moe/gpt_oss_triton_kernels_moe.py
index 5862abe20518..f50e162fb44c 100644
--- a/vllm/model_executor/layers/fused_moe/gpt_oss_triton_kernels_moe.py
+++ b/vllm/model_executor/layers/fused_moe/gpt_oss_triton_kernels_moe.py
@@ -552,9 +552,9 @@ def _supports_current_device() -> bool:
         cap = p.get_device_capability()
         if cap is None:
             return False
-        # (9,0) <= cap < (11,0) covers CUDA SM90 (Hopper), SM100+ (Blackwell)
-        # and ROCm gfx942/gfx950 (which map to 9.4/9.5).
-        return (9, 0) <= (cap.major, cap.minor) < (11, 0)
+        # (9,0) <= cap < (13,0) covers CUDA SM90 (Hopper), SM100+ (Blackwell)
+        # and ROCm gfx942/gfx950 (9.4/9.5) + gfx1200/gfx1201 (12.0).
+        return (9, 0) <= (cap.major, cap.minor) < (13, 0)
 
     @staticmethod
     def _supports_no_act_and_mul() -> bool:
@@ -884,9 +884,9 @@ def _supports_current_device() -> bool:
         cap = p.get_device_capability()
         if cap is None:
             return False
-        # (9,0) <= cap < (11,0) covers CUDA SM90 (Hopper), SM100+ (Blackwell)
-        # and ROCm gfx942/gfx950 (which map to 9.4/9.5).
-        return (9, 0) <= (cap.major, cap.minor) < (11, 0)
+        # (9,0) <= cap < (13,0) covers CUDA SM90 (Hopper), SM100+ (Blackwell)
+        # and ROCm gfx942/gfx950 (9.4/9.5) + gfx1200/gfx1201 (12.0).
+        return (9, 0) <= (cap.major, cap.minor) < (13, 0)
 
     @staticmethod
     def _supports_no_act_and_mul() -> bool:
diff --git a/vllm/model_executor/layers/fused_moe/oracle/mxfp4.py b/vllm/model_executor/layers/fused_moe/oracle/mxfp4.py
index ddc6588dc517..39c946dcbdd7 100644
--- a/vllm/model_executor/layers/fused_moe/oracle/mxfp4.py
+++ b/vllm/model_executor/layers/fused_moe/oracle/mxfp4.py
@@ -202,7 +202,7 @@ def select_mxfp4_moe_backend(
     triton_kernels_supported = has_triton_kernels() and (
         9,
         0,
-    ) <= current_platform.get_device_capability() < (11, 0)
+    ) <= current_platform.get_device_capability() < (13, 0)
 
     # LoRA: separate experts backend path
     if config.is_lora_enabled: